If you look upon the rationality principle from the point of view which I have here adopted, then you will find that it has little or nothing to do with the empirical or psychological assertion that man always, or in the main, or in most cases, acts rationally. Rather, it turns out to be an aspect of, or a consequence of, the methodological postulate that we should pack or cram our whole theoretical effort, our whole explanatory theory, into an analysis of the situation: into the model.

If we adopt this methodological postulate, then, as a consequence, the animating law will become a kind of zero principle. For the principle may be stated in this way: having constructed our model, our situation, we assume no more than that the actors act

4

within the terms of the model, or that they ‘work out’ what was implicit in the situation. This, incidentally, is what the term ‘situational logic’ alludes to.

The adoption of the rationality principle can therefore be regarded as a byproduct of a methodological postulate. It does not play the role of an empirical explanatory theory, of a testable hypothesis. For in this field, the empirical explanatory theories or hypotheses are our various models, our various situational analyses. It is these which may be empirically more or less adequate; which may be discussed and criticized, and whose adequacy may sometimes even be tested. And it is our analysis of a concrete empirical situation which may fail some empirical test, thereby enabling us to learn from our mistakes.

Tests of a model, it has to be admitted, are not easily obtainable and usually not very clearcut. But this difficulty arises even in the physical sciences. It is connected, of course, with the fact that models are always and necessarily rough; that they are always and necessarily schematic over simplifications. Their roughness entails a comparatively low degree of testability; for it is difficult to decide what is a discrepancy due to the necessary roughness, and what is a discrepancy indicative of a failure, a refutation of the model. Nevertheless, we can sometimes decide by tests which one of two (or more) competing models is the best. And in the social sciences, tests of a situational analysis can sometimes be provided by historical research.

But if the rationality principle does not play the role of an empirical or psychological proposition, and more especially, if it is not treated as subject on its own to any kind of tests: if tests, when available, are used to test a particular model, a particular situational analysis, of which the rationality principle forms a part; then even if a test decides that a certain model is less adequate than another one, since both operate with the rationality principle, we have no occasion to test this principle.

This remark explains, I think, why the rationality principle has been frequently declared to be a priori valid. And indeed, if it is not empirically refutable what else could it be but a priori valid?

The point is of considerable interest. Those who say that the rationality principle is a priori mean, of course, that it is a priori valid, or a priori true. But it seems to me quite clear that they must be mistaken. For the rationality principle seems to me clearly false - even in its weakest zero formulation which may be put like this: ‘Agents always act in a manner appropriate to the situation in which they find themselves.’

I think one can see very easily that this is not so. One has only to observe a flustered driver, desperately trying to park his car when there is no parking space to be found, in order to see that we do not always act in accordance with the rationality principle. Moreover, there are obviously vast personal differences, not only in knowledge and skill - these are part of the situation - but in assessing or understanding a situation; and this means that some people will act appropriately and others not.

But a principle that is not universally true is false. Thus the rationality principle is false. I think there is no way out of this. Consequently I must deny that it is a priori valid.

Now if the rationality principle is false, then an explanation which consists of the conjunction of this principle and a model must also be false, even if the particular model in question is true.

But can the model be true? Can any model be true? I do not think so. Any model, whether in physics or in the social sciences, must be an oversimplification. It must omit much, and it must overemphasize much.

My views on the rationality principle have been closely questioned. I have been asked whether there is not some confusion in what I say about the status of the ‘principle of acting adequately to the situation’ (that is, of my own version of the ‘rationality principle’); I was told, quite rightly, that I should make up my mind whether I want it to be a methodological principle, or an empirical conjecture. In the first case it would be clear that, and why, it could not be empirically tested; also why it could not be empirically false (but only part of a successful or unsuccessful methodology). In the second case, it would become part of the various social theories - the animating part of every social model. But then it would have to be part of some empirical theory, and would have to be tested along with the rest of that theory, and rejected if found wanting.

This second case is precisely the one that corresponds to my own view of the status of the rationality principle: I regard the principle of adequacy of action (that is, the rationality principle) as an integral part of every, or nearly every, testable social theory.

Now if a theory is tested, and found faulty, then we have always to decide which of its various constituent parts we shall make accountable for its failure. My thesis is that it is sound methodological policy to decide not to make the rationality principle accountable but the rest of the theory; that is, the model.

In this way it may appear that in our search for better theories we treat the rationality principle as if it were a logical or a metaphysical principle exempt from refutation: as unfalsifiable, or as a priori valid. But this appearance is misleading. There are, as I have indicated, good reasons to believe that the rationality principle, even in my minimum formulation, is actually false, though a good approximation to the truth. Thus it cannot be said that I treat it as a priori valid.

I hold, however, that it is good policy, a good methodological device, to refrain from blaming the rationality principle for the breakdown of our theory: we learn more if we blame our situational model.

The main argument in favour of this policy is that our model is far more interesting and informative, and far better testable, than the principle of the adequacy of our actions. We do not learn much in learning that this is not strictly true: we know this already. Moreover, in spite of being false, it is as a rule sufficiently near to the truth: if we can refute our theory empirically, then its breakdown will, as a rule, be pretty drastic, and though the falsity of the rationality principle may be a contributing factor, the main responsibility will normally attach to the model. Another point is this: the attempt to replace the rationality principle by another one seems to lead to complete arbitrariness in our model-building. And we must not forget that we can test a theory only as a whole, and that the test consists in finding the better of two competing theories which may have much in common; and most of them have the rationality principle in common.

But did not Churchill say, in The World Crisis, that wars are not won but only lost - that, in effect, they are competitions in incompetence? And does not this remark provide us with a kind of model for typical social and historical situations; a kind of model which emphatically is not animated by the rationality principle of the adequacy of our actions, but by a principle of inadequacy ?

The answer is that Churchill’s dictum means that most war

leaders are inadequate to their task, that they do not see the situation as it is, rather than that their actions cannot be understood (in good approximation at least) as adequate for the situation as they see it.

In order to understand their (inadequate) actions, we have therefore to reconstruct a wider view of the situation than their own. This must be done in such a way that we can see how and why the situation as they saw it (with their limited experience, their limited or overblown aims, their limited or overexcited imagination) led them to act as they did; that is to say, adequately for their inadequate view of the situational structure. Churchill himself uses this method of interpretation with great success, for example in his careful analysis of the failure of the Auchinleck/Ritchie team (in volume IV of The Second World War).

It is interesting to see that we employ the rationality principle to the limit of what is possible whenever we try to understand an action, even the action of a madman. We try to explain a madman’s actions, as far as possible, by his aims (which may be monomaniac) and by the ‘information’ on which he acts, that is to say, by his convictions (which may be obsessions, that is, false theories so tenaciously held that they become practically incorrigible). In so explaining the actions of a madman we explain them in terms of our wider knowledge of a problem situation which comprises his own, narrower, view of his problem situation; and understanding his actions means seeing their adequacy according to his view - his madly mistaken view - of the problem situation.

We may in this way even try to explain how he arrived at his madly mistaken view: how certain experiences shattered his originally sane view of the world and led him to adopt another -the most rational view he could develop in accordance with the information at his disposal, so far as he found it credible; and how he had to make this new view incorrigible, precisely because it would break down at once under the pressure of refuting instances which would leave him (so far as he could see) stranded without any interpretation of his world: a situation to be avoided at all costs, from a rational point of view, since it would make all rational action impossible.

Freud has often been described as the discoverer of human irrationality; but this is a misinterpretation, and a very superficial

one to boot. Freud’s theory of the typical origin of a neurosis falls entirely into our scheme: of explanations with the help of a situational model plus the rationality principle. For he explains a neurosis as an attitude adopted (in early childhood) because it was the best available way out of a situation which the agent (the child, the patient) was unable to understand and cope with. Thus the adoption of a neurosis becomes a rational act - as rational, say, as the act of a man who, jumping back when confronted by the danger of being run over by a car, collides with a bicyclist. It is rational

in the sense that the agent chose what appeared to him the immediately or obviously preferable or perhaps just the lesser evil - the less intolerable of two possibilities.

I shall say no more here about Freud’s method of therapy than that it is even more rationalistic than his method of diagnosis or explanation; for it is based on the assumption that once a man fully understands what befell him as a child, his neurosis will pass away.

But if we thus explain everything in terms of the rationality principle, does it not become tautological? By no means; for a tautology is obviously true, whilst we make use of the rationality principle merely as a good approximation to the truth, recognizing that it is not true, but false.

But if this is so, what becomes of the distinction between rationality and irrationality? Between mental health and mental disease?

This is an important question. The main distinction, I suggest, is that a healthy person’s beliefs are not incorrigible: a healthy person shows a certain readiness to correct his beliefs. He may do so only reluctantly, yet he is nevertheless ready to correct his views under the pressure of events, of the opinions held by others, and of critical arguments.

If this is so then we can say that the mentality of the man with definitely fixed views, the ‘committed’ man, is akin to that of the madman. It may be that all his fixed opinions are ‘adequate’ in the sense that they happen to coincide with the best opinion available at the time. But in so far as he is committed, he is not rational: he will resist any change, any correction; and since he cannot be in possession of the full truth (nobody is) he will resist rational

correction of even wildly mistaken beliefs. And he will resist, even if their correction is widely accepted during his lifetime.

Thus when those who praise commitment and irrational faith describe themselves as irrationalists (or post-rationalists) I agree with them. They are irrationalists, even if they are capable of reasoning. For they take pride in rendering themselves incapable of breaking out of their shell; they make themselves prisoners of their manias. They make themselves spiritually unfree, by an action whose adoption we may explain (following the psychiatrists) as one that is rationally understandable; understandable, for example, as an action they commit owing to fear - fear of being compelled, by criticism, to surrender a view which they dare not give up since they make it (or believe they must make it) the basis of their whole life. (Commitment - even ‘free commitment’ - and fanaticism, which, we know, can border on madness, are thus related in the most dangerous manner.)

To sum up: we should distinguish between rationality as a personal attitude (which, in principle, all sane men are capable of sharing) and the rationality principle.

Rationality as a personal attitude is the attitude of readiness to correct one’s beliefs. In its intellectually most highly developed form it is the readiness to discuss one’s beliefs critically, and to correct them in the light of critical discussions with other people.

The ‘rationality principle’ on the other hand has nothing to do with the assumption that men are rational in this sense - that they always adopt a rational attitude. It is, rather, a minimum principle (since it assumes no more than the adequacy of our actions to our problem situations as we see them) which animates all, or almost all, our explanatory situational models, and, although we know it not to be true, we have some reason to regard as a good approximation. Its adoption reduces considerably the arbitrariness of our models; an arbitrariness which becomes capricious indeed if we try to proceed without this principle.

Rationality, in the sense of an appeal to a universal and impersonal standard of truth, is of supreme importance ..., not only in ages in which it easily prevails, but also, and even more, in those less fortunate times in which it is despised and rejected as the vain dream of men who lack the virility to kill where they cannot agree.

BERTRAND RUSSELL¹

It can hardly be doubted that Hegel’s and Marx’s historicist philosophies are characteristic products of their time - a time of social change. Like the philosophies of Heraclitus and Plato, and like those of Comte and Mill, Lamarck and Darwin, they are philosophies of change, and they witness to the tremendous and undoubtedly somewhat terrifying impression made by a changing social environment on the minds of those who live in this environment. Plato reacted to this situation by attempting to arrest all change. The more modem social philosophers appear to react very differently, since they accept, and even welcome, change; yet this love of change seems to me a little ambivalent. For even though they have given up any hope of arresting change, as historicists they try to predict it, and thus to bring it under rational control; and this certainly looks like an attempt to tame it. Thus it seems that, to the historicist, change has not entirely lost its terrors.

In our time of still more rapid change, we even find the desire not only to predict change, but to control it by centralized large-scale planning. These holistic views (which I have criticized [in selection 24]) represent a compromise, as it were, between Platonic and Marxian theories. Plato’s will to arrest change, combined with Marx’s doctrine of its inevitability, yield, as a kind of Hegelian ‘synthesis’, the demand that since it cannot be entirely arrested, change should at least be ‘planned’, and controlled by the state, whose power is to be vastly extended.

An attitude like this may seem, at first sight, to be a kind of rationalism; it is closely related to Marx’s dream of the ‘realm of freedom’ in which man is for the first time master of his own fate. But as a matter of fact, it occurs in closest alliance with a doctrine which is definitely opposed to rationalism (and especially to the doctrine of the rational unity of mankind [see section n of selection 2]), one which is well in keeping with the irrationalist and mystical tendencies of our time. I have in mind the Marxist doctrine that our opinions, including our moral and scientific opinions, are determined by class interest, and more generally by the social and historical situation of our time. Under the name of ‘sociology of knowledge’ or ‘sociologism’, this doctrine has been developed recently (especially by M. Scheler and K. Mannheim²) as a theory of the social determination of scientific knowledge.

The sociology of knowledge argues that scientific thought, and especially thought on social and political matters, does not proceed in a vacuum, but in a socially conditioned atmosphere. It is influenced largely by unconscious or subconscious elements. These elements remain hidden from the thinker’s observing eye because they form, as it were, the very place which he inhabits, his social habitat. The social habitat of the thinker determines a whole system of opinions and theories which appear to him as unquestionably true or self-evident. They appear to him as if they were logically and trivially true, such as, for example, the sentence ‘All tables are tables’. This is why he is not even aware of having made any assumptions at all. But that he has made assumptions can be seen if we compare him with a thinker who lives in a very different social habitat; for he too will proceed from a system of apparently unquestionable assumptions, but from a very different one; and it may be so different that no intellectual bridge may exist and no compromise be possible between these two systems. Each of these different socially determined systems of assumptions is called by the sociologists of knowledge a total ideology.

The sociology of knowledge can be considered as a Hegelian version of Kant’s theory of knowledge. For it continues on the lines of Kant’s criticism of what we may term the ‘passivist’ theory of knowledge. I mean by this the theory of the empiricists down to and including Hume, a theory which may be described, roughly, as holding that knowledge streams into us through our senses, and that error is due to our interference with the sense-given material, or to the associations which have developed within it; the best way of avoiding error is to remain entirely passive and receptive. Against this receptacle theory of knowledge (I have called it the ‘bucket theory of the mind’ [see selection 7, section iv above]) Kant³ argued that knowledge is not a collection of gifts received by our senses and stored in the mind as if it were a museum, but that it is very largely the result of our own mental activity; that we must most actively engage ourselves in searching, comparing, unifying, generalizing, if we wish to attain knowledge. We may call this theory the ‘activist’ theory of knowledge. In connection with it, Kant gave up the untenable ideal of a science which is free from any kind of presuppositions. (That this ideal is even selfcontradictory was shown [in selection 2].) He made it quite clear that we cannot start from nothing, and that we have to approach our task equipped with a system of presuppositions which we hold without having tested them by the empirical methods of science; such a system may be called a ‘categorial apparatus’.⁴ Kant believed that it was possible to discover the one true and unchanging categorial apparatus, which represents as it were the necessarily unchanging framework of our intellectual outfit, i.e. human ‘reason’. This part of Kant’s theory was given up by Hegel, who, as opposed to Kant, did not believe in the unity of mankind. He taught that man’s intellectual outfit was constantly changing, and that it was part of his social heritage; accordingly the development of man’s reason must coincide with the historical development of his society, i.e. of the nation to which he belongs. This theory of Hegel’s, and especially his doctrine that all knowledge and all truth are ‘relative’ in the sense of being determined by history, is sometimes called ‘historism’ (in contradistinction to ‘historicism’). The sociology of knowledge or ‘sociologism’ is obviously very closely related to or nearly identical with it, the only difference being that, under the influence of Marx, it emphasizes that the historical development does not produce one uniform ‘national spirit’, as Hegel held, but rather several and sometimes opposed ‘total ideologies’ within one nation according to the class, the social stratum, or the social habitat, of those who hold them.

But the likeness to Hegel goes further. I have said above that according to the sociology of knowledge, no intellectual bridge or compromise between different total ideologies is possible. But this radical scepticism is not really meant quite as seriously as it sounds. There is a way out of it, and the way is analogous to the Hegelian method of superseding the conflicts which preceded him in the history of philosophy. Hegel, a spirit freely poised above the whirlpool of the dissenting philosophies, reduced them all to mere components of the highest of syntheses, of his own system. Similarly, the sociologists of knowledge hold that the ‘freely poised intelligence’ of an intelligentsia which is only loosely anchored in social traditions may be able to avoid the pitfalls of the total ideologies; that it may even be able to see through, and to unveil, the various total ideologies and the hidden motives and other determinants which inspire them. Thus the sociology of knowledge believes that the highest degree of objectivity can be reached by the freely poised intelligence analysing the various hidden ideologies and their anchorage in the unconscious. The way to true knowledge appears to be the unveiling of unconscious assumptions, a kind of psychotherapy, as it were, or if I may say so, a sociotherapy. Only he who has been socioanalysed or who has socioanalysed himself, and who is freed from this social complex, i.e. from his social ideology, can attain to the highest synthesis of objective knowledge.

In chapter 15 of The Open Society and Its Enemies, when dealing with ‘Vulgar Marxism’, I mentioned a tendency which can be observed in a group of modem philosophies, the tendency to unveil the hidden motives behind our actions. The sociology of knowledge belongs to this group, together with psychoanalysis and certain philosophies which unveil the ‘meaninglessness’ of the tenets of their opponents [see notes 17 and 18 to selection 6 above]. The popularity of these views lies, I believe, in the ease with which they can be applied, and in the satisfaction which they confer on those who see through things, and through the follies of the unenlightened. This pleasure would be harmless, were it not that all these ideas are liable to destroy the intellectual basis of any discussion, by establishing what I have called a ‘re-inforced dogmatism’. (Indeed, this is something rather similar to a ‘total ideology’.) Hegelianism does it by declaring the admissibility and even fertility of contradictions. But if contradictions need not be avoided, then any criticism and any discussion become impossible since criticism always consists in pointing out contradictions either within the theory to be criticized, or between it and some facts of experience. [See also selection 16 above.] The situation with psychoanalysis is similar: the psychoanalyst can always explain away any objections by showing that they are due to the repressions of the critic. And the philosophers of meaning, again, need only point out that what their opponents hold is meaningless, which will always be true, since ‘meaninglessness’ can be so defined that any discussion about it is by definition without meaning.⁵ Marxists, in a like manner, are accustomed to explain the disagreement of an opponent by his class bias, and the sociologists of knowledge by his total ideology. Such methods are both easy to handle and good fun for those who handle them. But they clearly destroy the basis of rational discussion, and they must lead, ultimately, to antirationalism and mysticism.

In spite of these dangers, I do not see why I should entirely forgo the fun of handling these methods. For just like the psychoanalysts, the people to whom psychoanalysis applies best,⁶ the socioanalysts invite the application of their own methods to themselves with an almost irresistible hospitality. For is not their description of an intelligentsia which is only loosely anchored in tradition a very neat description of their own social group? And is

it not also clear that, assuming the theory of total ideologies to be

.

correct, it would be part of every total ideology to believe that one’s own group was free from bias, and was indeed that body of the elect which alone was capable of objectivity? Is it not, therefore, to be expected, always assuming the truth of this theory, that those who hold it will unconsciously deceive themselves by producing an amendment to the theory in order to establish the objectivity of their own views? Can we, then, take seriously their claim that by their sociological self-analysis they have reached a higher degree of objectivity; and their claim that socioanalysis can cast out a total ideology? But we could even ask whether the whole theory is not simply the expression of the class interest of this particular group;

of an intelligentsia only loosely anchored in tradition, though just firmly enough to speak Hegelian as their mother tongue.

How little the sociologists of knowledge have succeeded in sociotherapy, that is to say, in eradicating their own total ideology, will be particularly obvious if we consider their relation to Hegel. For they have no idea that they are just repeating him; on the contrary, they believe not only that they have outgrown him, but also that they have successfully seen through him, socioanalysed him; and that they can now look at him, not from any particular social habitat, but objectively, from a superior elevation. This palpable failure in self-analysis tells us enough.

But, all joking apart, there are more serious objections. The sociology of knowledge is not only self-destructive, not only a rather gratifying object of socioanalysis, it also shows an astounding failure to understand precisely its main subject, the social aspects of knowledge, or rather, of scientific method. It looks upon science or knowledge as a process in the mind or ‘consciousness’ of the individual scientist, or perhaps as the product of such a process. If considered in this way, what we call scientific objectivity must indeed become completely ununderstandable, or even impossible; and not only in the social or political sciences where class interests and similar hidden motives may play a part but just as much in the natural sciences. Everyone who has an inkling of the history of the natural sciences is aware of the passionate tenacity which characterizes many of its quarrels. No amount of political partiality can influence political theories more strongly than the partiality shown by some natural scientists in favour of their intellectual offspring. If scientific objectivity were founded, as the sociologistic theory of knowledge naively assumes, upon the individual scientist’s impartiality or objectivity, then we should have to say goodbye to it. Indeed, we must be in a way more radically sceptical than the sociology of knowledge; for there is no doubt that we are all suffering under our own system of prejudices (or ‘total ideologies’, if this term is preferred); that we all take many things as self-evident, that we accept them uncritically and even with the naive and cocksure belief that criticism is quite unnecessary; and scientists are no exception to this rule, even though they may have superficially purged themselves from some of their prejudices in their particular field. But they have not

purged themselves by socioanalysis or any similar method; they have not attempted to climb to a higher plane from which they can understand, socioanalyse, and expurgate their ideological follies. For by making their minds more ‘objective’ they could not possibly attain to what we call ‘scientific objectivity’. No, what we usually mean by this term rests on different grounds.⁷ It is a matter of scientific method. And, ironically enough, objectivity is closely bound up with the social aspect of scientific method, with the fact that science and scientific objectivity do not (and cannot) result from the attempts of an individual scientist to be ‘objective’, but from the friendly-hostile co-operation of many scientists. Scientific objectivity can be described as the intersubjectivity of scientific method. But this social aspect of science is almost entirely neglected by those who call themselves sociologists of knowledge.

Two aspects of the method of the natural sciences are of importance in this connection. Together they constitute what I may term the ‘public character of scientific method’. First, there is something approaching free criticism. A scientist may offer his theory with the full conviction that it is unassailable. But this will not impress his fellow scientists and competitors; rather it challenges them: they know that the scientific attitude means criticizing everything, and they are little deterred even by authorities. Secondly, scientists try to avoid talking at cross purposes. (I may remind the reader that I am speaking of the natural sciences, but a part of modern economics may be included.) They try very seriously to speak one and the same language, even if they use different mother tongues. In the natural sciences this is achieved by recognizing experience as the impartial arbiter of their controversies. When speaking of‘experience’ I have in mind experience of a ‘public’ character, like observations, and experiments, as opposed to experience in the sense of more ‘private’ aesthetic or religious experience; and an experience is ‘public’ if everybody who takes the trouble can repeat it. In order to avoid speaking at cross purposes, scientists try to express their theories in such a form that they can be tested, i.e. refuted (or else corroborated) by such experience.

This is what constitutes scientific objectivity. Everyone who has learnt the technique of understanding and testing scientific theories can repeat the experiment and judge for himself. In spite of this, there will always be some who come to judgements which are partial, or even cranky. This cannot be helped, and it does not seriously disturb the working of the various social institutions which have been designed to further scientific objectivity and criticism; for instance the laboratories, the scientific periodicals, the congresses. This aspect of scientific method shows what can be achieved by institutions designed to make public control possible, and by the open expression of public opinion, even if this is limited to a circle of specialists. Only political power, when it is used to suppress free criticism, or when it fails to protect it, can impair the functioning of these institutions, on which all progress, scientific, technological, and political, ultimately depends.

In order to elucidate further still this sadly neglected aspect of scientific method, we may consider the idea that it is advisable to characterize science by its methods rather than by its results.

Let us first assume that a clairvoyant produces a book by dreaming it, or perhaps by automatic writing. Let us assume, further, that years later as a result of recent and revolutionary scientific discoveries a great scientist (who has never seen that book) produces one precisely the same. Or to put it differently, we assume that the clairvoyant ‘saw’ a scientific book which could not then have been produced by a scientist owing to the fact that many relevant discoveries were still unknown at that date. We now ask: is it advisable to say that the clairvoyant produced a scientific book? We may assume that, if submitted at the time to the judgement of competent scientists, it would have been described as partly ununderstandable, and partly fantastic; thus we shall have to say that the clairvoyant’s book was not when written a scientific work, since it was not the result of scientific method. I shall call such a result, which, though in agreement with some scientific results, is not the product of scientific method, a piece of ‘revealed science’.

In order to apply these considerations to the problem of the publicity of scientific method, let us assume that Robinson Crusoe succeeded in building on his island physical and chemical laboratories, astronomical observatories, etc., and in writing a great number of papers, based throughout on observation and experiment. Let us even assume that he had unlimited time at his disposal, and that he succeeded in constructing and in describing scientific systems which actually coincide with the results accepted at present by our own scientists. Considering the character of this Crusonian science, some people will be inclined, at first sight, to assert that it is real science and not ‘revealed science’. And, no doubt, it is very much more like science than the scientific book which was revealed to the clairvoyant, for Robinson Crusoe applied a good deal of scientific method. And yet, I assert that this Crusonian science is still of the ‘revealed’ kind; that there is an element of scientific method missing, and consequently, that the fact that Crusoe arrived at our results is nearly as accidental and miraculous as it was in the case of the clairvoyant. For there is nobody but himself to check his results; nobody but himself to correct those prejudices which are the unavoidable consequence of his peculiar mental history; nobody to help him to get rid of that strange blindness concerning the inherent possibilities of our own results which is a consequence of the fact that most of them are reached through comparatively irrelevant approaches. And concerning his scientific papers, it is only in attempts to explain his work to somebody who has not done it that he can acquire the discipline of clear and reasoned communication which too is part of scientific method. In one point - a comparatively unimportant one - is the ‘revealed’ character of the Crusonian science particularly obvious; I mean Crusoe’s discovery of his ‘personal equation’ (for we must assume that he made this discovery), of the characteristic personal reaction time affecting his astronomical observations. Of course it is conceivable that he discovered, say, changes in his reaction time, and that he was led, in this way, to make allowances for it. But if we compare this way of finding out about reaction time, with the way in which it was discovered in ‘public’ science - through the contradiction between the results of various observers - then the ‘revealed’ character of Robinson Crusoe’s science becomes manifest.

To sum up these considerations, it may be said that what we call ‘scientific objectivity’ is not a product of the individual scientist’s impartiality, but a product of the social or public character of scientific method; and the individual scientist’s impartiality is, so far as it exists, not the source but rather the result of this socially or institutionally organized objectivity of science.

Both⁸ Kantians and Hegelians make the same mistake of

assuming that our presuppositions (since they are, to start with, undoubtedly indispensable instruments which we need in our active ‘making’ of experiences) can neither be changed by decision nor refuted by experience; that they are above and beyond the scientific methods of testing theories, constituting as they do the basic presuppositions of all thought. But this is an exaggeration, based on a misunderstanding of the relations between theory and experience in science. It was one of the greatest achievements of our time when Einstein showed that, in the light of experience, we may question and revise our presuppositions regarding even space and time, ideas which had been held to be necessary presuppositions of all science, and to belong to its ‘categorial apparatus’. Thus the sceptical attack upon science launched by the sociology of knowledge breaks down in the light of scientific method. The empirical method has proved to be quite capable to taking care of itself.

But it does so not by eradicating our prejudices all at once; it can eliminate them only one by one. The classical case in point is again Einstein’s discovery of our prejudices regarding time. Einstein did not set out to discover prejudices; he did not even set out to criticize our conceptions of space and time. His problem was a concrete problem of physics, the redrafting of a theory that had broken down because of various experiments which in the light of the theory seemed to contradict one another. Einstein together with most physicists realized that this meant that the theory was false. And he found that if we alter it in a point which had so far been held by everybody to be self-evident and which had therefore escaped notice, then the difficulty could be removed. In other words, he just applied the methods of scientific criticism and of the invention and elimination of theories, of trial and error. But this method does not lead to the abandonment of all our prejudices; rather, we can discover the fact that we had a prejudice only after having got rid of it.

But it certainly has to be admitted that, at any given moment, our scientific theories will depend not only on the experiments, etc., made up to that moment, but also upon prejudices which are taken for granted, so that we have not become aware of them (although the application of certain logical methods may help us to detect them). At any rate, we can say in regard to this incrustation that science is capable of learning, of breaking down some of its crusts. The process may never be perfected, but there is no fixed barrier before which it must stop short. Any assumption can, in principle, be criticized. And that anybody may criticize constitutes scientific objectivity.

Scientific results are ‘relative’ (if this term is to be used at all) only in so far as they are the results of a certain stage of scientific development and liable to be superseded in the course of scientific progress. But this does not mean that truth is ‘relative’. If an assertion is true, it is true for ever. [See selection 14 above, especially sections i and il] It only means that most scientific results have the character of hypotheses, i.e. statements for which the evidence is inconclusive, and which are therefore liable to revision at any time. These considerations (with which I have dealt more fully elsewhere), though not necessary for a criticism of the sociologists, may perhaps help to further the understanding of their theories. They also throw some light, to come back to my main criticism, on the important role which co-operation, intersubjectivity, and the publicity of method play in scientific criticism and scientific progress.

It is true that the social sciences have not yet fully attained this publicity of method. This is due partly to the intelligence-destroying influence of Aristotle and Hegel, partly perhaps also to their failure to make use of the social instruments of scientific objectivity. Thus they are really ‘total ideologies’, or putting it differently, some social scientists are unable, and even unwilling, to speak a common language. But the reason is not class interest, and the cure is not a Hegelian dialectical synthesis, nor self-analysis. The only course open to the social sciences is to forget all about the verbal fireworks and to tackle the practical problems of our time with the help of the theoretical methods which are fundamentally the same in all sciences. I mean the methods of trial and error, of inventing hypotheses which can be practically tested, and of submitting them to practical tests. A social technology is needed whose results can be tested by piecemeal social engineering.

The cure here suggested for the social sciences is diametrically opposed to the one suggested by the sociology of knowledge. Sociologism believes that it is not their unpractical character, but rather the fact that practical and theoretical problems are too much intertwined in the field of social and political knowledge, that creates the methodological difficulties of these sciences. Thus we can read in a leading work on the sociology of knowledge:⁹ The peculiarity of political knowledge, as opposed to “exact” knowledge, lies in the fact that knowledge and will, or the rational element and the range of the irrational, are inseparably and essentially interwined.’ To this we can reply that ‘knowledge’ and ‘will’ are, in a certain sense, always inseparable; and that this fact need not lead to any dangerous entanglement. No scientist can know without making an effort, without taking an interest; and in his effort there is usually even a certain amount of self-interest involved. The engineer studies things mainly from a practical point of view. So does the farmer. Practice is not the enemy of theoretical knowledge but the most valuable incentive to it. Though a certain amount of aloofness may be becoming to the scientist, there are many examples to show that it is not always important for a scientist to be thus disinterested. But it is important for him to remain in touch with reality, with practice, for those who overlook it have to pay by lapsing into scholasticism. Practical application of our findings is thus the means by which we may eliminate irrationalism from social science, and not any attempt to separate knowledge from ‘will’.

As opposed to this, the sociology of knowledge hopes to reform the social sciences by making the social scientists aware of the social forces and ideologies which unconsciously beset them. But the main trouble about prejudices is that there is no such direct way of getting rid of them. For how shall we ever know that we have made any progress in our attempt to rid ourselves from prejudice? Is it not a common experience that those who are most convinced of having got rid of their prejudices are the most prejudiced? The idea that a sociological or a psychological or an anthropological or any other study of prejudices may help us to rid ourselves of them is quite mistaken; for many who pursue these studies are full of prejudice; and not only does self-analysis not help us to overcome the unconscious determination of our views, it often leads to even more subtle self-deception. Thus we can read in the same work on the sociology of knowledge¹⁰ the following references to its own activities: ‘There is an increasing tendency towards making conscious the factors by which we have so far been unconsciously

ruled____ Those who fear that our increasing knowledge of

determining factors may paralyse our decisions and threaten “freedom” should put their minds at rest. For only he is truly determined who does not know the most essential determining factors but acts immediately under the pressure of determinants unknown to him’. Now this is clearly just a repetition of a pet idea of Hegel’s which Engels naively repeated when he said: ‘Freedom is the appreciation of necessity.’¹¹ And it is a reactionary prejudice. For are those who act under the pressure of well-known determinants, for example, of a political tyranny, made free by their knowledge? Only Hegel could tell us such tales. But that the sociology of knowledge preserves this particular prejudice shows clearly enough that there is no possible shortcut to rid us of our ideologies. (Once a Hegelian, always a Hegelian.) Self-analysis is no substitute for those practical actions which are necessary for establishing the democratic institutions which alone can guarantee the freedom of critical thought, and the progress of science.